Abstract
OBJECTIVE: To assess the diagnostic and prognostic accuracy of C-Reactive Protein (CRP) in adult patients with suspected or confirmed sepsis. METHODS: A systematic review and meta-analysis were conducted using the Medline, Embase, and Google Scholar databases, without language or time restrictions. Studies involving adults at risk for, suspected of, or diagnosed with sepsis who underwent CRP measurement and reported diagnostic or prognostic accuracy data were eligible. Both cross-sectional and cohort studies were included. Risk of bias was evaluated with the QUADAS-2 tool, and the quality of evidence was graded using the GRADE approach. RESULTS: A total of 3599 records were screened, and 22 studies (13,083 patients) met the inclusion criteria, 16 assessing diagnostic accuracy and 6 evaluating prognostic value. In the diagnostic meta-analysis, the pooled sensitivity was 83 % (95 % CI 75 %-89 %) and specificity was 56 % (95 % CI 41 %-69 %), with high heterogeneity (I(2) = 80.1 %) and a false-positive rate of 44.3 % (95 % CI 30.9 %‒58.6 %). In the prognostic analysis for mortality, the pooled sensitivity was 81 % (95 % CI 70 %-89 %) and the specificity was 77 % (95 % CI 64 %-86 %), with high heterogeneity (I(2) = 65.9 %) and a false-positive rate of 22.8 % (95 % CI 13.6 %-35.6 %). Most studies presented a moderate to high risk of bias, and the overall quality of the evidence was rated as low. CONCLUSIONS: Current evidence reveals high diagnostic uncertainty and limited prognostic accuracy of CRP in adults with sepsis. The low specificity and high error rates restrict the clinical utility of CRP for diagnostic or prognostic decision-making in sepsis management.